Crystal and molecular structures of some phosphane-substituted cymantrenes [(C5H4 X)Mn(CO)LL′] (X = H or Cl, L = CO, L′ = PPh3 or PCy3, and LL’ = Ph2PCH2CH2PPh2)

The syntheses and structures of the cymantrenes [(C5H4 X)Mn(CO)LL′] (X = H or Cl; L = CO: L′ = PPh3 or PCy3; LL′ = Ph2PCH2CH2PPh2) are reported. Substitution of CO by phosphanes influences the bond parameters more than replacing the C5H5 ligand by C5H4Cl.


Introduction
The substitution of carbon monoxide (CO) by other donor ligands, particularly phosphanes, is one of the most important textbook examples for the reactivity of metal carbonyl complexes (Elschenbroich, 2016;Crabtree, 2005;Jordan, 2007). This is also true for the so-called 'piano-stool' complexes, which contain, besides CO ligands, aromatic -ligands like benzene or the cyclopentadienyl anion. Many studies have shown that the nature of the -ligand strongly influences the ease of CO substitution (Veiros, 2000;Le Moigne et al., 1976). But vice versa, the aromatic reactivity depends also on the electronic situation within the metal carbonyl moiety (Fan & Hall, 2001). One of the most studied systems is the 'cymantrene' series CpMn(CO) 3 and its substituted derivatives (Caulton, 1981). The substitution of one or two CO ligands by mono-or bidentate phosphanes was studied in the 1960s and it was found that the best way to do this was by UV irradiation (Strohmeier & Barbeau, 1962;Nyholm et al., 1963;Khatami et al., 1972a;Kursanov et al., 1970;Young & Wrighton, 1989). The choice of solvent and the irradiation time were the main determinants for the formation of either mono-or disubstitution products. Later on, studies on the spectroscopic [IR, ESR (electron spin resonance) and NMR] (Rehder & Keçeci, 1985;Ginzburg et al., 1974;Pike et al., 1989) and electro- ISSN 2053ISSN -2296 chemical properties (Treichel et al., 1975;Connelly & Kitchen, 1977) followed, which showed, as might have been expected, that the introduction of aryl-or alkylphosphanes led to increased electron density at the metal. Further studies were devoted to the reactivity in protonation reactions (Ginzburg et al., 1974), electrophilic hydrogen exchange reactions (Setkina et al., 1973;Khatami et al., 1972b;Antonova & Shapiro, 1991) and deprotonation by butyl lithium (Loim et al., 1988). A survey of the Cambridge Structural Database (CSD, Version 5.42, accessed on 26th August, 2021; Groom et al., 2016) showed no crystal structures for the fragments [(C 5 H 4 Cl)Mn-(CO)P] and about 80 entries for the corresponding C 5 H 5containing fragments. Limitation of the search to the fragment [(C 5 H 5 )Mn(CO) 2 PPh 2 ] gave 10 hits, of which most contained unsymmetrical mono-or dinuclear diphosphanes. Relevant in the context of this study were an early determination of the structure of [(C 5 H 5 )Mn(CO) 2 (PPh 3 )] ( Barbeau et al., 1972) and the crystal structure of [(C 5 H 5 )Mn(CO) 2 PPh 2 CH 2 Ph] (CSD refcode GIXRIO; Geicke et al., 1998). No hits were obtained for chelating diphosphanes, except for a derivative of 1,1 0 -bisdiphenylphosphanylferrocene (EFUHAO; André -Bentabet et al., 2002). We felt it might contribute to a better understanding of this substance class to add some more crystal structure determinations.
2.1.1. General procedure for the synthesis of 1a, 2a and 3a.
A solution of [(C 5 H 5 )Mn(CO) 3 ] (I) and a slight molar excess of the phosphane in tetrahydrofuran (THF, 120 ml) was irradiated for 7 h under argon. The colours of the solutions changed from yellow to red with concomitant gas evolution. After further stirring for 16 h, the solvent was evacuated and the residue dissolved in diethyl ether (Et 2 O) and filtered through a plug of silica gel. The solvent was evaporated again and the residue dissolved in the minimum amount of petroleum ether. This solution was placed on top of a silica gel chromatography column (alumina in the case of 3a) and the products were eluted with a petroleum ether/Et 2 O (9:1 v/v) mixture. Evaporation of the eluate yielded the products as yellow powders. Recrystallization from petroleum ether (with some added Et 2 O) by slow evaporation in a refrigerator at 5 C yielded crystals of all three compounds.

Refinement
In the refinements of 2a and 2b, a rigid-body restraint was used for the C3-C4 and C2-C3 bonds, respectively, because they had failed the 'Hirshfeld-Test' of PLATON (Spek, 2020) significantly. All H atoms were constrained. For compound 3a, PLATON analysis showed 16% solvent-accessible voids. Therefore, the SQUEEZE program (Spek, 2015) was used, which recovered 221 e per unit cell. Crystal data, data collection and structure refinement details are summarized in Table 1.
For the discussion of hydrogen bonds, the standard settings of Mercury (Macrae et al., 2020) (H atoms present, D-HÁ Á ÁA angle > 120.0 , 'all donors', contact distance range 'sum of vdW radii minus 5.00 to sum of vdW radii plus 0.00') were used for all compounds except 1b, where the 'sum of vdW radii plus 0.10' was used as the upper limit.
Irradiation of THF solutions of [(C 5 H 4 X)Mn(CO) 3 ] in the presence of PPh 3 leads to 1a and 1b in moderate yields of 40-60% (Scheme 1). Substantial amounts of the starting materials could be recovered. Products were isolated by chromatography and recrystallized from petroleum ether/Et 2 O.
3.1.1. Molecular and crystal structure of 1a. The crystals of 1a obtained from petroleum ether/Et 2 O are apparently a different modification than those described in the literature. Our compound crystallized in the monoclinic space group P2 1 /n with two independent molecules in the asymmetric unit (Fig. 1).
The major difference between the two molecules is in the relative orientation of the Mn(CO) 2 P tripod and the projection of the cyclopentadienyl ring. While in molecule A both Mn!P and one Mn!CO vector nearly eclipse C-H bonds of the cyclopentadienyl ring, in molecule B this is the case for the Mn!P vector only. In addition, the Mn2-P2 bond [2.2421 (7) Å ] is significantly longer (>20) than the Mn1-P1 bond [2.2259 (6) Å ]. All other bond lengths are identical in the two molecules (Table 2).
There are five intermolecular C-HÁ Á ÁO hydrogen bonds (Table S1 in the supporting information). Three of them involve arene C-H bonds, and carbonyl atom O22 accepts two of them (Fig. S1).
3.1.2. Molecular structure of 1b. Compound 1b crystallizes in the acentric orthorhombic space group P2 1 2 1 2 1 with one molecule in the asymmetric unit (Fig. 2)  The molecular structure (side view) of compound 1b, with displacement ellipsoids drawn at the 30% probability level.

Figure 1
The molecular structures of (a) molecule A and (b) molecule B of compound 1a, with displacement ellipsoids drawn at the 30% probability level.
program PLATON (Spek, 2020) showed no extra crystallographic symmetry and no sign of racemic twinning. The only 'molecular' origin of chirality resides in the PPh 3 'propeller'.
The Mn!P vector is nearly perpendicular to the C-Cl bond (torsion angle C1-Ct-Mn-P1 is 77.6 ). The individual bond lengths are nearly identical to those in 1a; the largest deviation is found for the C-C bonds of the cyclopentadienyl ring, which are slightly (1.5) shorter in 1b. The most important bond parameters can be found in Table 2.
There is only one intramolecular C-HÁ Á ÁCl hydrogen bond with a length shorter than the sum of the van der Waals radii (H16Á Á ÁCl1). Additionally, there is one weak intramolecular and three intermolecular C-HÁ Á ÁO hydrogen bonds, and one intermolecular C-HÁ Á ÁCl hydrogen bond ( Fig. S2 and Table S1 in the supporting information). The Cl atoms always bridge two different H atoms of the same symmetry-related arene ring along the a screw axis. Apparently, this interaction enforces the orientation of this particular arene ring and generates the chirality.

[(C 5 H 4 X)Mn(CO) 2 (PCy 3 )], X = H (2a) and Cl (2b)
The tricyclohexylphosphane compound 2a was first described in 1967 (Strohmeier & Mü ller, 1967) as part of a study on the -acceptor strength of phosphane ligands. It was then characterized by IR spectroscopy and elemental analysis. Later on it was shown that its reactivity in hydrogen isotope exchange reactions was more than 15 times greater in comparison to the PPh 3 compound 1a (Setkina et al., 1973). Further spectroscopic characterizations ( 13 C and 31 P NMR) and protonation studies followed soon afterwards (Ginzburg The molecular structures (side view) of compounds 2a (left) and 2b (right), with displacement ellipsoids drawn at the 30% probability level. Table 2 Important bond parameters of 1-3 in comparison with two related literature compounds.
Ct is the centroid of the cyclopentadienyl ring, (C-C) av is the average C-C bond length within the cyclopentadienyl ring and C x -Ct-Mn-P is the smallest torsion angle involving a cyclopentadienyl C-H (1a, 2a and 3a) or C-Cl bond.  , 1974). The chlorocyclopentadienyl complex 2b has not been reported before. We prepared both compounds according to Scheme 1 via irradiation of the corresponding tricarbonyl complexes in the presence of PCy 3 (tricyclohexylphosphane) in very low yield. Despite long irradiation times, large amounts of the starting material could be recovered. In contrast to 1a, it was not possible to lithiate 2a with n-BuLi or t-BuLi and chlorinate the presumed intermediate lithium compound with C 2 Cl 6 to give 2b. It was possible, however, to obtain crystals of both compounds suitable for X-ray diffraction.
3.2.1. Molecular structure of 2a. Compound 2a crystallizes in the monoclinic space group P2 1 /n, with one molecule in the asymmetric unit (Fig. 3). The Mn!P vector nearly eclipses a C-H bond of the cyclopentadienyl ring. While the Mn-P bond [2.2661 (7) Å ] is significantly longer (50) than the average Mn-P bond in 1a, the Mn-CO bonds are slightly shorter (3-5) ( Table 2). The distance from manganese to the cyclopentadienyl centroid is slightly longer (3) in 2a compared to 1a.
There is one intramolecular and two intermolecular C-HÁ Á ÁO hydrogen bonds involving exclusively methylene H atoms of the PCy 3 ligand and carbonyl atom O1. A packing diagram shows that these interactions mainly (although not exclusively) join the individual molecules in the c direction ( Fig. S3 and Table S1 in the supporting information).
3.2.2. Molecular structure of 2b. Compound 2b crystallizes in the monoclinic space group P2 1 /c, with one molecule in the asymmetric unit (Fig. 3). The Mn!P vector is nearly perpendicular to the C-Cl bond (torsion angle C1-Ct-Mn1-P1 is 78 ), with the Mn-P bond [2.2743 (9) Å ] being significantly longer (8) than in 2a. The Mn-CO bonds are slightly longer (3) than in 2a and have the same lengths as in 1b. This is also true for the distance of the Mn atom from the centroid of the cyclopentadienyl ring. More bond parameters can be found in Table 2.
There are intramolecular C-HÁ Á ÁX interactions involving two methylene H atoms of the PCy 3 ligand and either the Cl atom or one carbonyl O atom. Additionally, an intermolecular C-HÁ Á ÁCl hydrogen bond joins glide-plane-related molecules along the b axis ( Fig. S4 and Table S1 in the supporting information).
Irradiation of THF solutions of the corresponding tricarbonyl complexes in the presence of dppe for 7 h yields 3a and 3b in modest yields (30-40%), again with substantial recovery of the starting material. Some weak signals in the NMR spectra showed small amounts of other products, most likely dinuclear ones. However, the influence of prolonged reaction times on product yields and distribution was not examined. In contrast to the reactivity of 1b, it was not possible to deprotonate 3b [either by lithium diisopropylamide (LDA), lithium tetramethylpiperidide (LiTMP) or t-BuLi] and introduce more chlorine substituents via addition of C 2 Cl 6 . However, again it was possible to obtain crystals suitable for X-ray diffraction for both compounds.
3.3.1. Molecular structure of 3a. Compound 3a crystallizes in the monoclinic space group C2/c, with one molecule in the asymmetric unit. Fig. 4 shows a top view of the molecular structure. Both Mn!P vectors nearly eclipse two C-H bonds in mutual 1-and 3-positions of the cyclopentadienyl ring, while the Mn!CO vector bisects a C-C bond. The Mn-P [2.1968 (4) and 2.1849 (4) Å ] and Mn-CO [1.7549 (15) Å ] bonds, as well as the distance from manganese to the cyclopentadienyl centroid [1.761 (2) Å ], are shorter than for all the above-mentioned compounds. At the same time, the C-C The molecular structures (top views) of compounds 3a (left) and 3b (right), with displacement ellipsoids drawn at the 30% probability level. bonds of the cyclopentadienyl rings are longer than in the other compounds (Table 2).
There are two intermolecular hydrogen bonds involving the carbonyl O atom and one methylene H atom of the PCy 3 ligand or one C-H group of the cyclopentadienyl ring. The packing diagram shows that these interactions connect the individual molecules in the a direction ( Fig. S5 and Table S1 in the supporting information).
3.3.2. Molecular structure of 3b. Compound 3b crystallizes in the triclinic space group P1, with one molecule in the asymmetric unit (Fig. 4). The Mn!P2 vector bisects the C-C bond next to the chlorine substituent, while the Mn!P1 and Mn!CO vectors nearly eclipse two C-H bonds in the 2-and 4-positions. The Mn-P bond lengths [2.1961 (5) and 2.2024 (5) Å ] are significantly different from each other (by 12) and slightly longer than in 3a. The same holds for the relative distances between manganese and the cyclopentadienyl centroids, while the Mn-CO bonds are virtually identical (Table 2). It is worth noting the near perpendicular orientation of one arene ring (C201-C206) relative to the cyclopentadienyl ring (interplanar angle 86.0 ). This leads to a rather close approach of arene H atom H206 to cyclopentadienyl atom H4 (2.375 Å ).
There is one intermolecular C-HÁ Á ÁCl hydrogen bond involving an arene H atom, which joins the individual molecules in the b direction. The carbonyl O atom joins two molecules in the a direction, employing one arene H atom and one cyclopentadienyl H atom each ( Fig. S6 and Table S1 in the supporting information).

Comparison of the structures and conclusion
The introduction of a chlorine substituent in the cyclopentadienyl ring leads to a slight increase in the Mn-Ct and Mn-P distances for all the title phosphanes, while both the Mn-CO and the C-O bonds are only affected in the PCy 3 system, where a substantial elongation occurs. When comparing the two triads with different phosphanes, the Mn-Ct (Ct describes the centroid of the cyclopentadienyl ring) and Mn-P distances show a slight increase in the order 3!1!2. The C-O bonds follow the trend 1 ' 2 < 3 and the C-Cl bonds follow the trend 2b < 1b ' 3b. The average C-C bond lengths are the same within 2 for all six compounds. Comparison with the PPh 2 CH 2 Ph compound GIXRIO and the ferrocenylbisphosphane chelate compound EFUHAO shows more similarities with the PPh 3 complexes 1 than with the dppe chelates 3. The tendency of the Mn-P bonds to eclipse one cyclopentadienyl C-H bond is obvious in all the compounds. In all the chloro compounds, the Mn-P bonds avoid being eclipsed with the C-Cl bond of the cyclopentadienyl ring.
Apparently, the introduction of one chlorine substituent has only a small influence on the bond lengths, despite the relatively large effect on the spectroscopic data. Steric hindrance within the phosphanes seems to be of greater importance for the bond parameters than the differences in electronic effects. However, the presence of chlorine in the cyclopentadienyl ring leads to additional lattice stabilization via the formation of C-HÁ Á ÁCl hydrogen bonds.

Dicarbonyl(η 5 -cyclopentadienyl)(triphenylphosphane-κP)manganese(I) (compd1a)
Crystal data Hydrogen site location: inferred from neighbouring sites H-atom parameters constrained where P = (F o 2 + 2F c 2 )/3 (Δ/σ) max = 0.001 Δρ max = 0.55 e Å −3 Δρ min = −0.65 e Å −3 Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.

Special details
Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.

Fractional atomic coordinates and isotropic or equivalent isotropic displacement parameters (Å 2 )
x y z U iso */U eq C1 0.7611 ( where P = (F o 2 + 2F c 2 )/3 (Δ/σ) max < 0.001 Δρ max = 0.73 e Å −3 Δρ min = −1.63 e Å −3 Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes. Special details Geometry. All esds (except the esd in the dihedral angle between two l.s. planes) are estimated using the full covariance matrix. The cell esds are taken into account individually in the estimation of esds in distances, angles and torsion angles; correlations between esds in cell parameters are only used when they are defined by crystal symmetry. An approximate (isotropic) treatment of cell esds is used for estimating esds involving l.s. planes.